The expression data on lncRNA TMPO-AS1 in 33 types of human cancers were obtained from the Gene Expression Display Server GEDS (http://bioinfo.life.hust.edu.cn/web/GEDS/) (18). The expression of TMPO-AS1 in BC tissues and in normal tissues was analyzed using The Cancer Genome Atlas (TCGA) BLCA RNA-seq data retrieved from the UCSC XENA (https://xena.ucsc.edu), TANRIC (https://ibl.mdanderson.org/tanric/_design/basic/analysis.html) (19), and Gene Expression Omnibus (GEO, www.ncbi.nlm.nih.gov, GSE133624 and GSE120736 datasets) databases. The prognostic value of TMPO-AS1 was evaluated using GEPIA 2 (http://gepia2.cancer-pku.cn/#index). Additionally, hTFtarget (http://bioinfo.life.hust.edu.cn/hTFtarget#!/) and ChIPBase v2.0 (http://rna.sysu.edu.cn/chipbase/index.php) were used to find out the potential TFs of TMPO-AS1 (20, 21). The JASPAR (http://jaspar.genereg.net) 2018 database was used to identify the E2F1-TMPO-AS1 binding profile (22). Genes co-expressed with TMPO-AS1 (TCGA-BLCA dataset) were defined as those with the correlation coefficients ≥0.6 and p-values <0.01 using Co-LncRNA (http://bio-bigdata.hrbmu.edu.cn/Co-LncRNA/). (23). A pathway enrichment analysis was conducted using Metascape (https://metascape.org/gp/index.html#/main/step1) (24). The interactions between E2F1, OTUB1, and TMPO-AS1 were predicted via Agostini et al. introduced catRAPID (http://s.tartaglialab.com/page/catrapid_group), a server to identify the interaction of RNA and protein (25) and Tuvshinjargal et al. developed a web server named PRIdictor (http://bclab.inha.ac.kr/pridictor) to reveal mutual binding in protein and RNA (26). Additionally, the protein interactions between E2F1 and OTUB1 were predicted using HDOCK (http://hdock.phys.hust.edu.cn) (27). Last but not least, the ubiquitination sites of E2F1 were predicted using UbPred (http://www.ubpred.org) (28).

Resected BC and normal adjacent specimens were collected from patients with BC admitted to the Third Xiangya Hospital, Central South University, Hunan, China, from 2016 to 2018; all patients provided written informed consent. Six pairs of BC and paired adjacent normal tissues were stored in liquid nitrogen at −80°C. This study was approved by the ethics committee of the Third Xiangya Hospital, Central South University, Hunan, China.

The human BC cell lines, namely 5637, T24, and RT4, were obtained from American Type Culture Collection (ATCC; Rockville, MD, USA). BIU87 and EJ were purchased from the Advanced Research Center of Central South University (Changsha, China). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Invitrogen, Carlsbad, CA, USA) containing 10% fetal bovine serum (FBS; Gibco, Thermo Fisher Scientific, Waltham, MA, USA), 1 mmol/L glutamine, and 100 U/ml penicillin at 37°C in an incubator with 5% CO₂. The protein synthesis inhibitor cycloheximide (10 μg/ml, C1998; Millipore, Sigma-Aldrich, St Louis, MO, USA) and proteasome inhibitor MG132 (20 μM, S2619; Selleck, Houston, TX, USA) were used to examine the ubiquitin proteasome-related protein degradation.

Total RNA was extracted from BC tumor tissues, the paired adjacent normal tissues, and BC cells (T24 and RT4) using TRIzol (Invitrogen, Carlsbad, CA, USA) according to the instructions in the PrimeScript RT Reagent Kit (TaKaRa, Dalian, China). The quantitative real-time PCR (qRT-PCR) was performed using the SYBR Green PCR Master Mix (Toyobo, Osaka, Japan) as per the instructions of the manufacturer. The relative expression of genes was determined using the 2^−ΔΔCT method, and the expression was normalized to that of β-actin. All experiments were performed in triplicate. The primer sequences used in this study are listed in Table 1.

For in vitro functional assays, TMPO-AS1, E2F1, and OTUB1 overexpression plasmids and small interfering RNAs, as well as the empty vectors, were designed by GeneChem Co., Ltd (Shanghai, China) and transfected into T24 and RT4 cells using the Lipofectamine 3000 Reagent (Invitrogen, Carlsbad, CA, USA). For in vivo xenograft experiments, RT4 cells were stably transfected with empty lentiviral vectors, sh-TMPO-AS1 (designed according to the sequence of si-TMPO-AS1) or sh-TMPO-AS1, together with E2F1 overexpressing lentiviral vectors purchased from GeneChem Co., Ltd. (Shanghai, China) according to the protocol of the manufacturer. The empty vectors were used as the negative control. The transfection efficiency was determined via qRT-PCR.

The methyl thiazolyl tetrazolium (MTT) assay was conducted according to the instructions of the manufacturer. Briefly, the BC cells were plated into 96-well plates at a density of 1 × 10⁴ cells/well and incubated for 24 h. Later, 10 μl of the MTT Solution (Sigma Chemicals, St. Louis, MO, USA) was added to each well, and the cells were cultured at 37°C for 4 h. Furthermore, cell viability/proliferation was estimated via the measurement of the absorbance at 570 nm with the Epoch Microplate Spectrophotometer (BioTek Instruments Inc., Winooski, VT, USA).

The colony formation assay was performed as previously described in a study of Zeng et al. (29). Briefly, T24 and RT4 cells, treated as described in the abovementioned study, were seeded into 6-well plates at a density of 1,000 cells/well and cultured for 2–3 weeks. Then, the cells were washed with FBS, fixed with 4% paraformaldehyde, stained with 1% crystal violet, and counted. Only colonies with more than 50 cells were considered.

The cell apoptosis in BC cells (T24 and RT4 cells) was investigated via flow cytometry using the Annexin V-PE/7-AAD Kit (KA3809; Abnova, Wuhan, China) according to the protocol of the manufacturer.

The wound healing assay was conducted as previously described (30). Briefly, cells were seeded into 6-well plates at a density of 1 × 10⁵ cells/well. Then, a sterile 200 μl pipette tip was used to scratch a straight line in the cell monolayer. Later, the cells were washed with FBS and cultured for 48 h. The scratch width was measured 48 h later.

The transwell assay was conducted to evaluate cell migration and invasion. Briefly, 5 × 10⁴ BC cells (T24 and RT4) were suspended in the serum-free medium and seeded into the upper chamber of the Transwell Plates (8 μm pore; Corning, Corning, NY, USA) with Matrigel (BD Biosciences, San Jose, CA, USA), while the complete medium with 10% FBS was added to the lower chamber. After a 48-h incubation, the migrating cells were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet (Sigma-Aldrich, St. Louis, MO, USA), and photographed under a microscope (DMB5-2231P1, DMB HK Ltd., Hong Kong, China).

Total proteins were extracted using the radioimmunoprecipitation assay (RIPA) buffer (Beyotime Biotechnology Inc., Shanghai, China) with the Protease Inhibitor Cocktail (Roche, Basel, Switzerland). The protein concentration was measured using a bicinchoninic acid (BCA) kit (Thermo Fisher Scientific, Waltham, MA, USA). The protein samples were resolved via sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes. Later, the membranes were blocked in phosphate-buffered saline (PBS) containing 5% skim milk powder at room temperature for 1 h and incubated with the primary antibodies at 4°C overnight, followed by the secondary antibodies. The protein bands were visualized using the Pierce^® ECL Western Blotting Substrate Kit (32106; Thermo Fisher Scientific, Waltham, MA, USA) and normalized to the levels of β-actin as reference. The antibodies used in this study are listed in Supplementary Table 1.

Wild-type or mutant sequences of the E2F1 binding sites for the promoter of TMPO-AS1 were synthesized and inserted into the pGL3 vector (Promega, Madison, WI, USA). T24 and RT4 cells were seeded into 48-well plates and cotransfected with the above vectors along with E2F1 expression or control plasmids. About 48 h later, the luciferase activity was measured and analyzed using the Luciferase Reporter Assay System (Promega, Madison, WI, USA).

The chromatin immunoprecipitation (ChIP) was performed using the EZ Magna ChIP™ Kit (Millipore, Burlington, MA, USA) according to the instructions of the manufacturer. Briefly, 1 × 10⁷ cells (T24 and RT4) were fixed with 1% formaldehyde and treated with 10% glycine. Later, the cross-linked chromatin was broken into small DNA fragments via sonication. The sonicated DNA was immunoprecipitated using antibodies against E2F1 or control rabbit immunoglobulin G (IgG) (Bioss Antibodies Inc., Woburn, MA, USA). qRT-PCR was performed to quantify the precipitated chromatin using the specific primers listed in Table 1.

TMPO-AS1 was transcribed in vitro and labeled via 3′-end biotinylation. The RNA pull-down assay was performed using the Pierce™ Magnetic RNA-Protein Pull-Down Kit (Thermo Fisher Scientific, Waltham, MA, USA). Briefly, the lysates of control or TMPO-AS1 overexpressing T24 and RT4 cells were incubated with control or biotinylated TMPO-AS1 at room temperature for 4 h, followed by the addition of streptavidin magnetic beads (Thermo Fisher Scientific, Waltham, MA, USA) at 4°C for 60 min with rotation. After three washing steps with washing buffer, the RNA-binding proteins were eluted using 50 μl elution buffer and analyzed via Western blotting.

RNA immunoprecipitation was performed using the EZ-Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (Millipore, Burlington, MA, USA) based on the instructions of the manufacturer. Briefly, cell extracts were incubated with magnetic beads conjugated with antibodies against SNRNP70 (Cat.# CS203216) and anti-E2F1 (Cat.# OM250777) or with normal rabbit IgG (Cat.# PP64B). Anti-SNRNP70 and normal rabbit IgG antibodies were used as positive and negative controls, respectively. The relative abundance of TMPO-AS1 was normalized to the amount of enriched U1snRNA via qRT-PCR.

Cell lysates were incubated with primary antibodies against E2F1 (1:80, OM250777; Omnimabs, Alhambra, CA, USA) at 4°C overnight. Rabbit IgG antibodies (1:150, Bioss Antibodies Inc., Woburn, MA, USA) were used as the negative control. Later, the cell lysates were mixed with the protein A/G agarose (Cat.# P1012, Beyotime Biotechnology Inc., Jiangsu, China) at 4°C for 2 h, followed by centrifugation and washing steps. The precipitated complex was separated using the SDS-PAGE and analyzed via a Western blotting.

Fluorescence in situ hybridization (FISH) was performed using the Ribo™ Fluorescent in situ Hybridization Kit (RiboBio Co. Ltd., Guangzhou, China) following the instructions of the manufacturer. The TMPO-AS1 and 18S probes were synthesized and labeled with the Cy3 fluorescent dye. Fluorescence was detected under a Confocal Laser Microscope (SP5; Leica Microsystems, Wetzlar, Germany).

The tissue sections obtained from paraffin-embedded tissues were dewaxed in xylene and rehydrated in an ethanol gradient. Later, tissues were incubated in 1% hydrogen peroxide and boiled in citrate buffer (10 mM, pH = 6.0) for 15 min. Subsequently, tissues were incubated with the primary antibodies against Ki-67 (1:1,000, 27309-1-AP; Proteintech, Chicago, IL, USA), E2F1 (1:200, OM250777; Omnimabs, Alhambra, CA, USA), and caspase-3 (1:200, 19677-1-AP; Proteintech, Chicago, IL, USA) at 4°C overnight, followed by incubation with horseradish peroxidase (HRP)-conjugated goat anti-rabbit secondary antibody (SP-9000, zsbio, Beijing, China) at room temperature for 30 min. Diaminobenzidine was used as chromogen; hematoxylin was used as the nuclear counterstain.

All animal experiments were approved by the Animal Care and Use Committee of the Central South University. The 1 × 10⁶ RT4 cells transfected with empty vector, sh-TMPO-AS1, or sh-TMPO-AS1 together with E2F1-expressing lentiviral vectors were injected subcutaneously into the flanks of 4- to 6-week-old male BALB/c nude mice (n = 4 per group) obtained from the Shanghai Experimental Laboratory Animal Center (Shanghai, China). Tumor volumes were measured every 3 days and calculated as follows: tumor volume = (D × d²)/2, where D and d refer to the longest and shorter diameters, respectively. Mice were euthanized after 25 days.

All statistical analyses were performed using the GraphPad Prism Software, Version 8 (GraphPad Software, San Diego, CA, USA). Data are presented as the mean ± SD of at least three independent experiments. The relationship between E2F1 and TMPO-AS1 was analyzed using the Pearson's correlation coefficient. Significant differences were analyzed using the Student's t-test or the one-way ANOVA. Values of p < 0.05 were considered statistically significant.

To evaluate the expression of TMPO-AS1 in tumor and normal tissues, we used the online database GEDS; interestingly, we found that TMPO-AS1 is upregulated in multiple tumor tissues vs. normal tissues (Figure 1A). Consistently, the expression of TMPO-AS1 was upregulated in BC tissues compared with the normal tissues according to TCGA (Figure 1B) and TANRIC databases (Figure 1C). We further confirmed that the expression of TMPO-AS1 was higher in six BC tissues compared with that in the corresponding normal tissues via qRT-PCR (Figure 1D). Moreover, the integrative analysis of GSE133624 and GSE120736 showed that TMPO-AS1 was not only highly expressed in BC tissues (Figure 1E) but also exhibited higher levels in MIBC vs. NMIBC samples (Figure 1F). Of note, the higher expression of TMPO-AS1 was associated with the recurrence of BC (Figure 1G) and the advanced tumor stage (Supplementary Figure 1A). Furthermore, patients with BC of higher TMPO-AS1 expression levels were associated with shorter disease-free survival times (Figure 1H). Taken together, these data suggest that TMPO-AS1 is highly expressed in BC tissues, and it may serve as a potential prognostic biomarker in patients with BC.

Since the expression of TMPO-AS1 was the highest in T24 and RT4 cells among five BC cell lines (BIU87, 5637, T24, EJ, and RT4) as per qRT-PCR (Figure 2A), these two cell lines were selected for the following experiments. To investigate the effects of TMPO-AS1 on the proliferation, migration, and apoptosis of BC cells, loss- and gain-of-function assays were performed. First, si-RNA targeting TMPO-AS1 or a TMPO-AS1-overexpression plasmid was transfected into T24 and RT4 cells. The transfection efficacy was examined using qRT-PCR (Figure 2B). Importantly, TMPO-AS1 knockdown significantly inhibited cell viability/proliferation (Figures 2C,D), migration (Figure 2E), and invasion (Figure 2F), whereas the overexpression of TMPO-AS1 resulted in the opposite effects. Additionally, cell apoptosis was induced in TMPO-AS1-silenced T24 and RT4 cells, whereas the overexpression of TMPO-AS1 inhibited cell apoptosis (Figure 2G). Overall, these findings suggest that TMPO-AS1 plays an oncogenic role in BC cells, promoting their proliferation, migration, invasion, and survival in vitro.

To figure out the underlying mechanism of TMPO-AS1-mediated carcinogenesis in BC, we investigated the upstream and downstream targets of TMPO-AS1. We screened out the potential TFs of TMPO-AS1 using hTFtarget and ChIPBase v2.0 (Figure 3A). Interestingly, we obtained 36 candidate TFs for TMPO-AS1; of note, E2F1 had the highest positive correlation with TMPO-AS1 among the 36 TFs (Supplementary Table 2). Importantly, a previous study reported that the overexpression of E2F3 induced the promoter activity of TMPO-AS1/LAP2, an antisense transcript (31). Additionally, we analyzed the genes co-expressed with TMPO-AS1 (correlation coefficient ≥0.6 and p < 0.01) in BC using Co-LncRNA (Supplementary Table 3) and used them in the context of pathway analysis via Metascape. Remarkably, we found that TMPO-AS1 is involved in the E2F pathway (Figure 3B). Thus, we focused on E2F1 in the following experiments, and we examine its transcriptional expression among five BC cell lines (Supplementary Figure 1B). Importantly, E2F1 was positively correlated with TMPO-AS1 in BC as evidenced by TCGA and GSE133624 datasets (Figures 3C,D). Furthermore, the expression of TMPO-AS1 was downregulated after E2F1 silencing and upregulated in E2F1-overexpressing BC cells (Figure 3E). Of note, the putative binding site between TMPO-AS1 and E2F1 was located at around −945 to −935 bp, upstream of the transcription start site, as predicted by JASPAR (http://jaspar.genereg.net), a collection of transcription factor binding site profiles (22) (Figure 3F). Importantly, this prediction was validated via luciferase reporter assays. The overexpression of E2F1 dramatically enhanced the luciferase activity of the wild-type TMPO-AS1 promoter but did not affect the transcriptional activity of the mutant TMPO-AS1 promoter (Figure 3G). Moreover, the results of ChIP assays demonstrated that E2F1 was remarkably enriched in the TMPO-AS1 promoter region relative to the observed in the context of control IgG (Figure 3H). Overall, these results indicate that E2F1 binds to the promoter of TMPO-AS1 and its transcription.

Given that lncRNAs have been reported to interact with E2F1 (32), we investigated whether TMPO-AS1 could interact with E2F1. As shown in Figure 4A, a possible interaction between TMPO-AS1 and E2F1 was predicted by catRAPID. Consequently, we conducted RNA pull-down assays to validate the prediction. The results showed that E2F1 was abundantly enriched in the context of the TMPO-AS1 probe compared with the oligo control, especially in TMPO-AS1 overexpressing RT4 and T24 cells (Figure 4B). Similarly, RIP results demonstrated that E2F1 remarkably immunoprecipitated TMPO-AS1, particularly after the TMPO-AS1 overexpression in RT4 and T24 cells (Figure 4C). The interaction between TMPO-AS1 and E2F1 prompted us to further investigate whether TMPO-AS1 influenced the expression of E2F1. Interestingly, the overexpression or knockdown of TMPO-AS1 had no significant impact on the E2F1 mRNA expression (Figure 4D), but it either increased or decreased the protein levels of E2F1 in RT4 and T24 cells (Figure 4E), respectively. These results suggest that the TMPO-AS1-mediated E2F1 regulation occurs at the posttranscriptional level. Of note, we found that the proteasome inhibitor MG132 markedly increased the stability of E2F1 in cycloheximide-treated RT4 and T24 cells (Figure 4F), implying that the turnover of E2F1 is dependent on the ubiquitin-proteasome system. Importantly, TMPO-AS1 silencing significantly shortened the half-life of E2F1 (Figure 4G). Altogether, our data demonstrate that TMPO-AS1 directly interacts with E2F1 and regulates its protein levels via protein stabilization.

Later, we aimed to understand the underlying mechanism of TMPO-AS1-mediated E2F1 stabilization. First, using FISH, we found that TMPO-AS1 is predominantly distributed in the cytoplasm in both T24 and RT4 cells (Figure 5A), indicating that TMPO-AS1 might be involved in translational regulation. Of note, it was predicted that E2F1 has four potential ubiquitination sites, as per UbPred (Supplementary Figure 2A). Therefore, we hypothesized that TMPO-AS1 may associate with a specific ubiquitinase/deubiquitinase to regulate E2F1 ubiquitination. Since we have previously studied a group of deubiquitinases (OTUB1, UCHL5, USP5, COPS6, and PSMD14) in BC (unpublished data), we evaluated their binding potential to TMPO-AS1 using PRIdictor. The results highlighted OTUB1 as the deubiquitinase that most likely associates with TMPO-AS1 (Supplementary Figure 2B), and the transcriptional expression of OTUB1 was evaluated by qRT-PCR (Supplementary Figure 1C). To validate this prediction, we conducted RNA pull-down assays and found that TMPO-AS1 precipitated more OTUB1 than the other deubiquitinases (Figure 5B). Similarly, RIP assays showed that TMPO-AS1 was obviously immunoprecipitated by the anti-OTUB1 antibody (vs. IgG; Figure 5C). Furthermore, catRAPID showed that a region of TMPO-AS1 (located at 76–127 nt) exhibits a high potential of interaction with some of the OTUB1 amino acid residues (51–152; Supplementary Figure 2C), further supporting the association between TMPO-AS1 and OTUB1. Of note, despite the direct association between TMPO-AS1 and OTUB1, TMPO-AS1 failed to alter the OTUB1 protein expression (Figure 5D). Importantly, the OTUB1 knockdown led to a decrease in the E2F1 protein levels in RT4 and T24 cells (Figure 5E), suggesting an interaction between these two proteins. Such interaction was further supported by the prediction by HDOCK; OTUB1 is very likely to bind to E2F1 (Figure 5F). Importantly, the results of co-immunoprecipitation (Co-IP) assays showed that silencing TMPO-AS1 not only increased E2F1 ubiquitination but also mitigated the interaction between OTUB1 and E2F1 (Figure 5G); importantly, this phenotype was rescued after the TMPO-AS1 overexpression. Furthermore, the OTUB1 knockdown significantly reversed the decreased ubiquitination of E2F1 induced by the overexpression of TMPO-AS1 in RT4 and T24 cells (Figure 5H). Taken together, these results indicate that TMPO-AS1 upregulates E2F1 protein levels via OTUB1-mediated deubiquitination and the consequent protein stabilization.

E2F1 is associated with cell proliferation, apoptosis, metastasis, and invasiveness (33). Hence, we speculated that E2F1 would promote BC tumorigenesis and development. According to the data retrieved from TCGA and the GEO (GSE133624) databases, the expression of E2F1 was remarkably higher in BC tissues than that in normal tissues (Figures 6A,B). Furthermore, Western blotting and immunohistochemistry (IHC) analyses showed that the expression of E2F1 protein was upregulated in six paired BC tissues compared with the adjacent normal tissues (Figures 6C,D). As anticipated, further functional experiments demonstrated that the E2F1 knockdown significantly inhibited cell proliferation (Figures 6E,F), migration (Figure 6G), and invasion (Figure 6H), and induced apoptosis (Figure 6I), whereas the overexpression of E2F1 led to the opposite effects. Altogether, these results reveal that E2F1 promotes malignant phenotypes in BC cells.

Since the overexpression of either E2F1 or TMPO-AS1 enhanced the proliferation, migration, and invasion of BC cells, and TMPO-AS1 upregulated the protein levels of E2F1, we hypothesized that TMPO-AS1 would promote BC progression via E2F1. To prove this theory, we restored the expression of E2F1 in TMPO-AS1-silenced BC cells (Figure 7A) and performed MTT, colony formation, wound healing, and transwell assays. As expected, the restoration of E2F1 in BC cells significantly reversed the inhibition of cell proliferation (Figures 7B,C), migration (Figure 7D), and invasion (Figure 7E) induced by TMPO-AS1 silencing. Furthermore, the TMPO-AS1 silencing-mediated promotion of apoptosis could be abrogated by the E2F1 overexpression (Figure 7F). Therefore, our data demonstrate that TMPO-AS1 promotes cell proliferation, migration, invasion, and survival of BC via E2F1 in vitro.

Stably transfected RT4 cells were subcutaneously inoculated into nude mice to explore whether TMPO-AS1 would promote BC growth via E2F1 in vivo; every 5 days, the tumor volumes were measured. Results showed that the knockout of TMPO-AS1 significantly suppressed tumor growth compared to the control group, whereas the overexpression of E2F1 abolished the inhibitory effect on tumor growth induced by the knockout of TMPO-AS1 (Figures 8A,B). Furthermore, the IHC staining showed that the depletion of TMPO-AS led to a substantial decrease of the protein levels of Ki-67 and E2F1 and to a notable increase in the expression of caspase-3. Of note, this phenotype was reversed by the overexpression of E2F1 (Figure 8C). Collectively, the above findings demonstrate that TMPO-AS1 regulates the BC-associated tumor growth through E2F1 in vivo.